South Korea's Kospi index fell 3.1% to 6,692.61 following AI warnings. SK Hynix fell more than 5%, while Samsung dropped 4%.
Showing 1–10 of 100
Iranian Vice President for Nuclear Affairs Mohammad Eslami has been blocked from addressing the IAEA General Conference in Vienna. This occurred after the U.S. urged Austria to deny a sanctions waiver allowing his entry.
Google DeepMind launched AlphaGenome Atlas, a 1-petabyte AI-powered searchable database. It predicts the impact of all 9 billion possible single-letter DNA variants to help scientists understand disease and biology. The tool is freely available for academic research.
French AI startup Mistral has raised 3 billion euros after its latest funding. The firm is valued at 21 billion euros.
Anthropic said it halted potential plots by scientists who used its AI models for research that could have helped develop biological weapons. It also stated that accounts linked to governments tried to use its AI for potential bioweapons or public surveillance on multiple occasions.
Morgan Stanley and other banks have been adopting OpenAI's ChatGPT. OpenAI is launching a new tool, built with Morgan Stanley and Evercore, that uses GPT-6 Astra to research companies, analyze data, and generate pitchbooks to perform the work of junior bankers.
GPT-Live-1 is now available in the API, allowing anyone to DIY AI toys globally. The full-duplex voice model from OpenAI listens while speaking, replacing traditional three-stage systems, and is priced at $0.05 per minute for the front-end voice layer.
Anthropic announced Thursday that it banned accounts for potentially misusing its Claude AI model to develop biological weapons, revealing findings from an investigation. The company reported blocking potential AI-assisted bioweapons research by scientists.
A security researcher obtained a 6TB Fable dataset from a Chinese LLM router. The data contains SSH keys, VPN configs, Aliyun keys, and GitLab tokens, exposing seven Chinese and CIS government entities and 19 major firms, including Xiaomi, Huawei, NIO, and Minimax.
Anthropic released a 154-page threat intelligence report detailing attempts to misuse Claude for cyberattacks, biology, and building weapons. Anthropic disrupted every operation, including a case where a team created over 4,700 romance personas across 20 dating apps.